Sebastian Dröge [Tue, 1 Nov 2011 15:08:59 +0000 (16:08 +0100)]
omxvideodec: Implement draining of the component and use it
This makes sure that all buffers are decoded and pushed downstream
before flushing the ports and losing some buffers.
Sebastian Dröge [Thu, 20 Oct 2011 12:32:40 +0000 (14:32 +0200)]
omxvideodec: Fix deadlock if ::reset is called before finding the corresponding frame of a decoded buffer
Sebastian Dröge [Tue, 1 Nov 2011 14:10:12 +0000 (15:10 +0100)]
omxaudioenc: Forward downstream flow returns to upstream
Sebastian Dröge [Tue, 1 Nov 2011 14:10:01 +0000 (15:10 +0100)]
omxvideoenc: Forward downstream flow returns to upstream
Sebastian Dröge [Tue, 1 Nov 2011 12:58:38 +0000 (13:58 +0100)]
omxvideodec: Forward downstream flow returns to upstream
Sebastian Dröge [Tue, 25 Oct 2011 12:23:38 +0000 (14:23 +0200)]
omx: Add minimal README file
Sebastian Dröge [Thu, 20 Oct 2011 13:21:07 +0000 (15:21 +0200)]
omxvideodec: Fix deadlock caused by calling reset while the loop function does something with the base video codec stream lock
Sebastian Dröge [Thu, 20 Oct 2011 13:20:47 +0000 (15:20 +0200)]
omxvideoenc: Fix deadlock caused by calling reset while the loop function does something with the base video codec stream lock
Sebastian Dröge [Thu, 20 Oct 2011 12:30:38 +0000 (14:30 +0200)]
omxvideodec: Move locking at the correct place
Sebastian Dröge [Fri, 14 Oct 2011 08:27:47 +0000 (10:27 +0200)]
omx: Remove QCOM hack to reset nOffset in EmptyBufferDone
This is now done in a generic way that does not require any
hacks because it will work without any side effects on any
OMX implementation.
Sebastian Dröge [Fri, 14 Oct 2011 08:26:00 +0000 (10:26 +0200)]
omx: Reset nOffset to 0 in EmptyBufferDone if nFilledLen is 0
Some OMX implementations don't reset nOffset when the complete
buffer is emptied but instead only reset nFilledLen. We reset
nOffset to 0 if nFilledLen == 0, which is safe to do because
the offset *must* be 0 if the buffer is not filled at all.
Seen in QCOM's OMX implementation.
Sebastian Dröge [Tue, 4 Oct 2011 08:56:33 +0000 (10:56 +0200)]
omxvideoenc: If one parameter/configuration is not supported don't skip the next
Sebastian Dröge [Mon, 3 Oct 2011 07:12:44 +0000 (09:12 +0200)]
omxh264dec: Require stream-format=byte-stream
Other stream-formats are unlikely to be supported by OMX components.
Sebastian Dröge [Thu, 29 Sep 2011 08:37:32 +0000 (10:37 +0200)]
omxvideodec: Add API for subclasses to prepare/convert frames
Sebastian Dröge [Tue, 27 Sep 2011 13:08:54 +0000 (15:08 +0200)]
omxaudioenc: Switch to Idle first and wait before switching to Loaded and deallocating buffers
Allocating buffers before the Idle state is reached can lead to crashes.
Sebastian Dröge [Tue, 27 Sep 2011 13:08:44 +0000 (15:08 +0200)]
omxvideoenc: Switch to Idle first and wait before switching to Loaded and deallocating buffers
Allocating buffers before the Idle state is reached can lead to crashes.
Sebastian Dröge [Tue, 27 Sep 2011 13:05:19 +0000 (15:05 +0200)]
omxvideodec: Switch to Idle first and wait before switching to Loaded and deallocating buffers
Allocating buffers before the Idle state is reached can lead to crashes.
Sebastian Dröge [Tue, 27 Sep 2011 12:15:06 +0000 (14:15 +0200)]
omxvideodec: New hack for QCOM to recreate the component instead of reconfiguring it on caps changes
Sebastian Dröge [Tue, 27 Sep 2011 10:13:56 +0000 (12:13 +0200)]
omx: Change a g_assert() into a GST_WARNING_OBJECT()
Sebastian Dröge [Mon, 26 Sep 2011 11:04:18 +0000 (13:04 +0200)]
omx: Add hack for Qualcomm's OMX implementation to manually reset nOffset in EmptyBufferDone
Sebastian Dröge [Fri, 23 Sep 2011 15:02:49 +0000 (17:02 +0200)]
omxaudioenc: Add a check to prevent a zero-sized OMX buffer
Sebastian Dröge [Fri, 23 Sep 2011 15:02:19 +0000 (17:02 +0200)]
omxvideoenc: Add some more checks for correct OMX buffer sizes
Sebastian Dröge [Fri, 23 Sep 2011 13:53:49 +0000 (15:53 +0200)]
omxvideodec: Add some more checks for OMX buffer sizes
Sebastian Dröge [Wed, 14 Sep 2011 08:15:38 +0000 (10:15 +0200)]
omx: Wait until the Executing state is reached before calling OMX_FillThisBuffer()
This correctly works around the QCOM race condition that happens when calling
FTB after setting the new state and before reaching it.
Sebastian Dröge [Fri, 2 Sep 2011 12:43:43 +0000 (14:43 +0200)]
omxvideodec: Negotiate video format with downstream and what the component claims to support
Vincent Penquerc'h [Thu, 25 Aug 2011 18:56:58 +0000 (19:56 +0100)]
basevideoencoder: fix element leak
and this concludes an hour of yelling at the bloody test failing,
only to track down the problem not being in the test.
https://bugzilla.gnome.org/show_bug.cgi?id=657368
Sebastian Dröge [Fri, 19 Aug 2011 07:20:39 +0000 (09:20 +0200)]
omxvideoenc: Release basevideocodec stream lock while waiting for a buffer
This prevents deadlocks if no empty input buffers are available and
releasing input buffers requires the loop function to handle some
output buffers first.
Sebastian Dröge [Fri, 19 Aug 2011 07:19:22 +0000 (09:19 +0200)]
omxvideodec: Release basevideocodec stream lock while waiting for a buffer
This prevents deadlocks if no empty input buffers are available and
releasing input buffers requires the loop function to handle some
output buffers first.
Sebastian Dröge [Thu, 18 Aug 2011 08:24:26 +0000 (10:24 +0200)]
basevideodecoder: Fix deadlock
Sebastian Dröge [Thu, 18 Aug 2011 08:03:20 +0000 (10:03 +0200)]
baseaudiodecoder: Don't take the stream lock in the seek handler
This will lead to deadlocks
Sebastian Dröge [Thu, 18 Aug 2011 08:02:50 +0000 (10:02 +0200)]
basevideo: Fix locking, especially if both pads have different streaming threads
Sebastian Dröge [Thu, 18 Aug 2011 07:42:02 +0000 (09:42 +0200)]
basevideo: Don't call g_type_class_peek_parent() in class_init
This is already done by the GObject boilerplate macro
Sebastian Dröge [Thu, 18 Aug 2011 07:40:46 +0000 (09:40 +0200)]
baseaudiodecoder: Don't call g_type_class_peek_parent() in class_init
This is already done by the boilerplate macro
Sebastian Dröge [Thu, 18 Aug 2011 07:34:38 +0000 (09:34 +0200)]
baseaudiodecoder: Fix thread safety issues if both pads have different streaming threads
Sebastian Dröge [Thu, 18 Aug 2011 07:17:04 +0000 (09:17 +0200)]
baseaudiodecoder: Delay sending of serialized events to finish_frame()
Sebastian Dröge [Wed, 17 Aug 2011 12:33:31 +0000 (14:33 +0200)]
omxaudioenc: Always require at least OMX_MIN_PCMPAYLOAD_MSEC per input buffer
Sebastian Dröge [Wed, 17 Aug 2011 12:28:44 +0000 (14:28 +0200)]
baseaudioencoder: Add support for requesting a minimum and maximum number of samples per frame
This extends the special case of a fixed number of samples per frame
that was supported before already.
Sebastian Dröge [Wed, 17 Aug 2011 12:17:18 +0000 (14:17 +0200)]
omxaudioenc: Handle inbuf==NULL properly in ::handle_frame()
Sebastian Dröge [Wed, 17 Aug 2011 11:04:19 +0000 (13:04 +0200)]
omxaacenc: Implement ::get_num_samples() vfunc
Sebastian Dröge [Wed, 17 Aug 2011 11:03:50 +0000 (13:03 +0200)]
omxaudioenc: Add vfunc to get the number of samples inside a buffer
Sebastian Dröge [Wed, 17 Aug 2011 09:34:31 +0000 (11:34 +0200)]
omxaudioenc: Release baseaudioencoder stream lock while waiting for a buffer in ::handle_frame()
This prevents deadlocks if no empty input buffers are available and
releasing input buffers requires the loop function to handle some
output buffers first.
Sebastian Dröge [Wed, 17 Aug 2011 09:34:04 +0000 (11:34 +0200)]
baseaudioencoder: Fix thread safety issues if both pads have different streaming threads
Sebastian Dröge [Wed, 17 Aug 2011 07:58:01 +0000 (09:58 +0200)]
baseaudioencoder: Taking the OBJECT lock in reset() is not needed
Sebastian Dröge [Tue, 16 Aug 2011 09:03:24 +0000 (11:03 +0200)]
omxaudioenc: Remove hack that only applies to the video encoder class
Sebastian Dröge [Tue, 16 Aug 2011 08:49:21 +0000 (10:49 +0200)]
omxaacenc: Add initial version of OpenMAX AAC encoder element
Sebastian Dröge [Mon, 15 Aug 2011 13:10:04 +0000 (15:10 +0200)]
omxaudioenc: Add initial version of audio encoder base class
Sebastian Dröge [Mon, 15 Aug 2011 12:14:11 +0000 (14:14 +0200)]
baseaudioencoder: Delay sending of serialized events to finish_frame()
Sebastian Dröge [Mon, 15 Aug 2011 11:06:51 +0000 (13:06 +0200)]
audio: Integrate audio base classes into the build system and fixup
Sebastian Dröge [Mon, 15 Aug 2011 10:56:00 +0000 (12:56 +0200)]
audio: Add audio decoder/encoder base classes
Taken from http://cgit.collabora.com/git/user/manauw/gst-plugins-bad.git/log/?h=baseaudio
Sebastian Dröge [Fri, 12 Aug 2011 10:25:03 +0000 (12:25 +0200)]
basevideoencoder: Proxy the width/height/framerate/PAR constraints of downstream caps to upstream
This allows to specify constraints on the compressed downstream caps
by muxers or capsfilters, which will then be forwarded to upstream
and allows video converters to fulfill the constraints.
Code based on Mark Nauwelaerts audio encoder base class.
Sebastian Dröge [Fri, 12 Aug 2011 10:13:45 +0000 (12:13 +0200)]
basevideoencoder: Remove old ::getcaps() comment
Sebastian Dröge [Fri, 12 Aug 2011 10:06:23 +0000 (12:06 +0200)]
basevideoencoder: Remove ::get_caps() vfunc
Subclasses can set the caps more efficiently and this only
caused additional indirections.
Sebastian Dröge [Wed, 10 Aug 2011 08:24:21 +0000 (10:24 +0200)]
omxvideoenc: Use "video/x-raw-yuv" as sink template caps instead of strict I420 caps
Sebastian Dröge [Wed, 10 Aug 2011 08:23:39 +0000 (10:23 +0200)]
omxmpeg4videodec: Don't require width/height on sink pad caps
Sebastian Dröge [Wed, 10 Aug 2011 08:11:37 +0000 (10:11 +0200)]
omxvideodec: Use "video/x-raw-yuv" as src template caps instead of strict I420 caps
Sebastian Dröge [Wed, 10 Aug 2011 07:56:30 +0000 (09:56 +0200)]
omxvideoenc: Set the state back to StateLoaded even if an error happened
Sebastian Dröge [Wed, 10 Aug 2011 07:49:57 +0000 (09:49 +0200)]
omx: Don't hold any locks while calling OMX_SendCommand()
It might call into one of the callbacks and lead to deadlocks, e.g.
with the Qualcomm OMX implementation.
Sebastian Dröge [Wed, 10 Aug 2011 07:32:01 +0000 (09:32 +0200)]
omx: Move some code
Sebastian Dröge [Wed, 10 Aug 2011 07:23:10 +0000 (09:23 +0200)]
omx: Reset pending reconfigure output ports when changing the state from Executing to any lower state
Sebastian Dröge [Wed, 10 Aug 2011 07:08:00 +0000 (09:08 +0200)]
omx: Fix crash when setting last error after the ports were freed
Sebastian Dröge [Wed, 10 Aug 2011 07:03:52 +0000 (09:03 +0200)]
omx: Free component structure
Sebastian Dröge [Wed, 10 Aug 2011 07:02:52 +0000 (09:02 +0200)]
omx: Make component destruction safer
Sebastian Dröge [Wed, 10 Aug 2011 06:53:05 +0000 (08:53 +0200)]
omx: Set pAppPrivate of buffers to NULL when deallocating buffers
This prevents usage of freed memory later if the OMX component
has weird behaviour.
Sebastian Dröge [Wed, 10 Aug 2011 06:52:25 +0000 (08:52 +0200)]
omxvideodec: Set the state back to StateLoaded even if an error happened
Sebastian Dröge [Wed, 10 Aug 2011 06:51:54 +0000 (08:51 +0200)]
omx: Add some assertions to check if the buffer pAppPrivate is still correct
Sebastian Dröge [Mon, 8 Aug 2011 11:04:30 +0000 (13:04 +0200)]
omx: Add parenthesis at correct places in the struct init macro
Sebastian Dröge [Mon, 8 Aug 2011 10:12:58 +0000 (12:12 +0200)]
omx: Only prevent setting a higher state if the component is in an error state
Sebastian Dröge [Wed, 3 Aug 2011 14:02:01 +0000 (16:02 +0200)]
basevideodecoder: Use the cached video frame size instead of recalculating it
Sebastian Dröge [Wed, 3 Aug 2011 13:35:01 +0000 (15:35 +0200)]
omx: Improve debugging in param/config getter/setter wrappers
Sebastian Dröge [Wed, 3 Aug 2011 11:10:33 +0000 (13:10 +0200)]
omxvideodec: Don't abort if the color format is not supported but give a useful error message
Sebastian Dröge [Tue, 2 Aug 2011 13:14:37 +0000 (15:14 +0200)]
omxvideoenc: Don't fail if setting the bitrate or profile is not supported by the component
Also always set/get the profile, even if there are no peer caps.
Sebastian Dröge [Tue, 2 Aug 2011 13:14:24 +0000 (15:14 +0200)]
basevideoencoder: Make access to the list of frames threadsafe
Sebastian Dröge [Mon, 1 Aug 2011 11:22:05 +0000 (13:22 +0200)]
omx: Add a hacks flag for the Qualcomm 7x30 OMX_FillThisBuffer() race and make it optional
Sebastian Dröge [Fri, 29 Jul 2011 11:56:59 +0000 (13:56 +0200)]
omx: Add workaround for QCOM 7x30 race condition
Sebastian Dröge [Fri, 29 Jul 2011 10:06:21 +0000 (12:06 +0200)]
omxh263enc: Add H.263 encoder element
Sebastian Dröge [Fri, 29 Jul 2011 09:26:39 +0000 (11:26 +0200)]
omxmpeg4videoenc: Add support for setting profile/level via caps
Sebastian Dröge [Thu, 28 Jul 2011 12:14:45 +0000 (14:14 +0200)]
omxh264enc: Add support for setting profile/level via caps
Sebastian Dröge [Thu, 28 Jul 2011 10:58:25 +0000 (12:58 +0200)]
omxvideoenc: Add support for forcing the next frame to be a keyframe
Sebastian Dröge [Thu, 28 Jul 2011 09:54:16 +0000 (11:54 +0200)]
omxvideoenc: Add support for setting bitrate/quantization related parameters
Sebastian Dröge [Thu, 28 Jul 2011 08:23:08 +0000 (10:23 +0200)]
omx: Add wrapper functions for OMX_[GS]et{Config,Parameter}
Sebastian Dröge [Thu, 28 Jul 2011 07:54:53 +0000 (09:54 +0200)]
omx: Add macro to initialize OpenMAX structures
Sebastian Dröge [Thu, 28 Jul 2011 07:08:38 +0000 (09:08 +0200)]
omxvideoenc: Don't output 0-byte buffers
Sebastian Dröge [Mon, 25 Jul 2011 13:05:08 +0000 (15:05 +0200)]
omx: Ensure that the pAppPrivate pointer in OMX buffers is set correctly
Sebastian Dröge [Mon, 25 Jul 2011 11:19:06 +0000 (13:19 +0200)]
omxvideo{enc,dec}: Only set/unset flushing state on ports if they were created already
Sebastian Dröge [Mon, 25 Jul 2011 10:01:05 +0000 (12:01 +0200)]
omxwmvdec: Add WMV video decoder element
Sebastian Dröge [Mon, 25 Jul 2011 09:44:56 +0000 (11:44 +0200)]
omxh263dec: Add H.263 decoder element
Sebastian Dröge [Mon, 25 Jul 2011 09:32:51 +0000 (11:32 +0200)]
omxh264enc: Add H.264 encoder element
Sebastian Dröge [Mon, 25 Jul 2011 08:48:58 +0000 (10:48 +0200)]
omxvideodec: Try harder to deallocate the buffers after errors happened
Sebastian Dröge [Mon, 25 Jul 2011 08:47:28 +0000 (10:47 +0200)]
omxvideoenc: Try harder to deallocate the buffers after errors happened
Sebastian Dröge [Mon, 25 Jul 2011 08:46:49 +0000 (10:46 +0200)]
omx: Deallocate port buffers before freeing the component
They should be deallocated by the caller before reaching the
Loaded state but to be on the safe side we will make sure
they're really deallocated here.
Sebastian Dröge [Thu, 21 Jul 2011 09:15:14 +0000 (11:15 +0200)]
omxvideoenc: Add initial support for stride conversion
Sebastian Dröge [Thu, 21 Jul 2011 08:38:26 +0000 (10:38 +0200)]
omx: Set default roles for the components if none were set from the config file
Sebastian Dröge [Thu, 21 Jul 2011 08:36:19 +0000 (10:36 +0200)]
omx: Failure to set the component role is fatal
Sebastian Dröge [Thu, 21 Jul 2011 05:53:25 +0000 (07:53 +0200)]
omxvideoenc: Add support for setting codec_data on the srcpad caps
Sebastian Dröge [Thu, 21 Jul 2011 05:44:34 +0000 (07:44 +0200)]
omxvideoenc: Free/drop GstVideoFrames that resulted in an empty buffer
Sebastian Dröge [Thu, 21 Jul 2011 05:44:10 +0000 (07:44 +0200)]
basevideoencoder: Allow finishing of frames with no src_buffer to drop/free the GstVideoFrame
Sebastian Dröge [Thu, 21 Jul 2011 05:31:05 +0000 (07:31 +0200)]
omxvideoenc: Remove obsolete TODO comment
Sebastian Dröge [Wed, 20 Jul 2011 09:09:54 +0000 (11:09 +0200)]
omx: Use libgstopenmax.so for the plugin filename and openmax for the plugin name
Resolves conflicts with gst-openmax.
Sebastian Dröge [Wed, 20 Jul 2011 06:34:33 +0000 (08:34 +0200)]
omxvideoenc: Add video encoder base class and MPEG4 video encoder
Unfortunately requires lots of hacks again to work properly with
Bellagio.
Sebastian Dröge [Wed, 20 Jul 2011 08:39:51 +0000 (10:39 +0200)]
basevideoencoder: Only get caps from the subclass if they were not set yet by the subclass